Page 410 - RUICHSS 2023 Proceeding

University of Ruhuna ISSN: 2706-0063
Matara, Sri Lanka

One latent variable's intermediate value was discovered in the data because of a mismatch

between the indicated number of latent variables to be extracted based on parallel analysis
and the eigenvalue test. Four intentions-related items made up the intention component,

which explained 79.7% of the variation in the data.

Figure 1: Scree Plot of the 28 Items

Source: Authors’ calculation, 2023

On a screen graph, the component number and eigenvalue are shown. Four components were

the only ones extracted for analysis after looking at Figure 01: screen plot. This would allow
us to propose that the variation in the data may be explained by four major components. On

their respective constructions, each indicator (factor) substantially loaded over 0.50 ( <

0.001). They were maintained in the measurement model because they had large factor
loadings. These results confirmed the robust uni-dimensionality and convergent validity of

the measurement model (Hair, et al., 2010).

Additionally, Hair Jr., Black et al. suggested that the reliability test be conducted before
starting the construct validity study and that the constructs are considered reliable when

Cronbach's alpha is 70 or above (Hair, et al., 2010).

351